Related Books of Interest

Enterprise Master The Business of IT Data Management How to Improve Service and Lower Costs An SOA Approach to Managing By Robert Ryan and Tim Raducha-Grace Core Information ISBN: 0-13-700061-8 By Allen Dreibelbis, Eberhard Hechler, Ivan Drive More Business Value from IT…and Bridge the Milman, Martin Oberhofer, Paul Van Run, and Gap Between IT and Business Leadership Dan Wolfson ISBN: 0-13-236625-8 IT organizations have achieved outstanding The Only Complete Technical Primer for MDM technological maturity, but many have been slower Planners, Architects, and Implementers to adopt world-class business practices. This book provides IT and business executives with methods Enterprise Master Data Management provides to achieve greater business discipline throughout an authoritative, vendor-independent MDM IT, collaborate more effectively, sharpen focus technical reference for practitioners: archi- on the customer, and drive greater value from IT tects, technical analysts, consultants, solution investment. Drawing on their experience consult- designers, and senior IT decision makers. ing with leading IT organizations, Robert Ryan and Written by the IBM® data management Tim Raducha-Grace help IT leaders make sense innovators who are pioneering MDM, this of alternative ways to improve IT service and lower book systematically introduces MDM’s key cost, including ITIL, IT financial management, concepts and technical themes, explains its balanced scorecards, and business cases. You’ll business case, and illuminates how it inter- learn how to choose the best approaches to relates with and enables SOA. improve IT business practices for your environment and use these practices to improve service quality, Drawing on their experience with cutting-edge reduce costs, and drive top-line revenue growth. projects, the authors introduce MDM patterns, blueprints, solutions, and best practices published nowhere else—everything you need to establish a consistent, manageable set of master data, and use it for competitive advantage.

Sign up for the monthly IBM Press newsletter at ibmpressbooks/newsletters Related Books of Interest

The Art of Enterprise Information Architecture The New Era of Enterprise A Systems-Based Approach for Business Intelligence: Unlocking Business Insight Using Analytics to Achieve a Global By Mario Godinez, Eberhard Hechler, Klaus Competitive Advantage Koenig, Steve Lockwood, Martin Oberhofer, By Mike Biere and Michael Schroeck ISBN: 0-13-707542-1 ISBN: 0-13-703571-3 A Complete Blueprint for Maximizing the Architecture for the Intelligent Enterprise: Value of Business Intelligence in the Powerful New Ways to Maximize the Enterprise Real-time Value of Information The typical enterprise recognizes the im- Tomorrow’s winning “Intelligent Enterprises” mense potential of business intelligence (BI) will bring together far more diverse sources and its impact upon many facets within the of data, analyze it in more powerful ways, and organization–but it’s not easy to transform deliver immediate insight to decision-makers BI’s potential into real business value. Top BI throughout the organization. Today, however, expert Mike Biere presents a complete blue- most companies fail to apply the information print for creating winning BI strategies and they already have, while struggling with the infrastructure, and systematically maximiz- complexity and costs of their existing ing the value of information throughout the information environments. enterprise.

In this book, a team of IBM’s leading This product-independent guide brings information management experts guide you together start-to-finish guidance and practical on a journey that will take you from where checklists for every senior IT executive, you are today toward becoming an “Intelligent planner, strategist, implementer, and the Enterprise.” actual business users themselves.

Listen to the author’s podcast at: ibmpressbooks.com/podcasts

Visit ibmpressbooks.com for all product information Related Books of Interest

Understanding DB2 9 Security Bond, See, Wong, Chan ISBN: 0-13-134590-7

DB2 9 for Linux, UNIX, and Windows Mining the Talk DBA Guide, Reference, and Unlocking the Business Value in Exam Prep, 6th Edition Unstructured Information Baklarz, Zikopoulos By Scott Spangler and Jeffrey Kreulen ISBN: 0-13-185514-X ISBN: 0-13-233953-6 Leverage Unstructured Data to Become More Competitive, Responsive, and Viral Data in SOA Innovative An Enterprise Pandemic Fishman In Mining the Talk, two leading-edge IBM ISBN: 0-13-700180-0 researchers introduce a revolutionary new approach to unlocking the busi- ness value hidden in virtually any form of unstructured data–from word processing IBM Cognos 10 Report documents to websites, emails to instant Studio messages. Practical Examples Draskovic, Johnson The authors review the business drivers ISBN-10: 0-13-265675-2 that have made unstructured data so important–and explain why conventional methods for working with it are inad- Data Integration Blueprint equate. Then, writing for business profes- sionals–not just data mining specialists– and Modeling they walk step-by-step through exploring Techniques for a Scalable and your unstructured data, understanding it, Sustainable Architecture and analyzing it effectively. Giordano ISBN-10: 0-13-708493-5

Sign up for the monthly IBM Press newsletter at ibmpressbooks/newsletters This page intentionally left blank Decision Management Systems This page intentionally left blank Decision Management Systems A Practical Guide to Using Business Rules and Predictive Analytics

James Taylor

IBM Press Pearson plc Upper Saddle River, NJ • Boston • Indianapolis • San Francisco New York • Toronto • Montreal • London • Munich • Paris • Madrid Cape Town • Sydney • Tokyo • Singapore • Mexico City ibmpressbooks.com The authors and publisher have taken care in the preparation of this book, but make no expressed or implied warranty of any kind and assume no responsibility for errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of the use of the information or programs contained herein. © Copyright 2012 by International Business Machines Corporation. All rights reserved. Note to U.S. Government Users: Documentation related to restricted right. Use, duplication, or disclosure is subject to restrictions set forth in GSA ADP Schedule Contract with IBM Corporation. IBM Press Program Managers: Steven M. Stansel, Ellice Uffer Cover design: IBM Corporation Associate Publisher: Dave Dusthimer Marketing Manager: Stephane Nakib Executive Editor: Mary Beth Ray Senior Development Editor: Kimberley Debus Managing Editor: Kristy Hart Designer: Alan Clements Technical Editors: Claye Greene, Don Griest Project Editor: Jovana San Nicolas-Shirley Indexer: Lisa Stumpf Compositor: Gloria Schurick Proofreader: Seth Kerney Manufacturing Buyer: Dan Uhrig Published by Pearson plc Publishing as IBM Press IBM Press offers excellent discounts on this book when ordered in quantity for bulk purchases or special sales, which may include electronic versions and/or custom covers and content particular to your business, training goals, marketing focus, and branding interests. For more information, please contact U. S. Corporate and Government Sales 1-800-382-3419 [email protected]. For sales outside the U. S., please contact International Sales [email protected]. The following terms are trademarks or registered trademarks of International Business Machines Corporation in the United States, other countries, or both: IBM, the IBM Press logo, SPSS, WebSphere, and ILOG. Netezza is a registered trademark of Netezza Corporation, an IBM Company. Microsoft is a trademark of Microsoft Corporation in the United States, other coun- tries, or both. Other company, product, or service names may be trademarks or service marks of others. The Library of Congress cataloging-in-publication data is on file. All rights reserved. This publication is protected by copyright, and permission must be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in any form or by any means, electronic, mechanical, photocopying, recording, or likewise. For information regarding permissions, write to: Pearson Education, Inc. Rights and Contracts Department 501 Boylston Street, Suite 900 Boston, MA 02116 Fax (617) 671-3447 ISBN-13: 978-0-13-288438-9 ISBN-10: 0-13-288438-0 Text printed in the United States on recycled paper at R.R. Donnelley in Crawfordsville, Indiana. First printing October 2011 For Meri, even though it’s not poetry For my parents And for my boys, again This page intentionally left blank Contents

Foreword by Deepak Advani xv Foreword by Pierre Haren xvii Preface xix Acknowledgments xxiii

Part I The Case for Decision Management Systems 1

Chapter 1 ■ Decision Management Systems Are Different 3 Agile 4 Analytic 8 Adaptive 15

Chapter 2 ■ Your Business Is Your Systems 19 Changing Expectations 20 Changing Scale 23 Changing Interactions 25

Chapter 3 ■ Decision Management Systems Transform Organizations 29 A Market of One 30 Always On 33 Breaking the Ratios 36 Crushing Fraud 39 Maximizing Assets 41 Maximizing Revenue 44 Making Smart People Smarter 45 Conclusion 46 xii Decision Management Systems

Chapter 4 ■ Principles of Decision Management Systems 47 Principle #1: Begin with the Decision in Mind 48 Principle #2: Be Transparent and Agile 57 Principle #3: Be Predictive, Not Reactive 60 Principle #4: Test, Learn, and Continuously Improve 63 Summary 67

Part II Building Decision Management Systems 69

Chapter 5 ■ Discover and Model Decisions 71 Characteristics of Suitable Decisions 72 A Decision Taxonomy 81 Finding Decisions 87 Documenting Decisions 99 Prioritizing Decisions 111

Chapter 6 ■ Design and Implement Decision Services 115 Build Decision Services 116 Integrate Decision Services 147 Best Practices for Decision Services Construction 152

Chapter 7 ■ Monitor and Improve Decisions 157 What Is Decision Analysis? 158 Monitor Decisions 159 Determine the Appropriate Response 167 Develop New Decision-Making Approaches 176 Confirm the Impact Is as Expected 184 Deploy the Change 187 Contents xiii

Part III Enablers for Decision Management Systems 189

Chapter 8 ■ People Enablers 191 The Three-Legged Stool 191 A Decision Management Center of Excellence 196 Organizational Change 206

Chapter 9 ■ Process Enablers 211 Managing a Decision Inventory 211 Adapting the Software Development Lifecycle 215 Decision Service Integration Patterns 221 A Culture of Experimentation 222 Moving to Fact-Based Decisioning 228 The OODA Loop 232

Chapter 10 ■ Technology Enablers 235 Business Rules Management Systems 235 Predictive Analytics Workbenches 238 Optimization Systems 243 Pre-Configured Decision Management Systems 244 Data Infrastructure 247 A Service-Oriented Platform 255 Epilogue 263 Bibliography 267 Index 273 This page intentionally left blank Foreword by Deepak Advani

Over the last couple of decades, businesses gained a competitive advantage by automating business processes. New companies and ecosystems were born around ERP, SCM, and CRM. We are at a point where automation is no longer a competitive advantage. The next wave of differentiation will come through decision optimization. And at the heart of decision optimization is a smart decision system, a topic that James Taylor does an outstanding job of explaining in this book. As James explains, a smart decision system encapsulates business rules, predictive models, and optimization. Business rules codify the best practices and human knowledge that a business builds up over time. Predictive models use statistics and mathematical algorithms to recommend the best action at any given time. Optimization, through constraint-based programming or mathematical programming tech- niques originally applied to operations research, delivers the best out- come. It is the combination of all three disciplines that enables organizations to optimize decisions. What used to be called artificial intelligence became predictive and advanced analytical techniques and are now Decision Management Systems, which are increasingly populat- ing business processes and making adopters competitive. As James describes in the book, a Decision Management System opti- mizes decisions not only for knowledge workers, but for all workers. This enables a call center representative to make the best offer to reduce customer churn, a claims processing worker to maximize fraud detec- tion, and a loan officer to reduce risk while maximizing return. And it’s not just decisions made by people— a Decision Management System can enable your e-commerce site to present the next best offer, traffic control systems to automatically make adjustments to reduce congestion, and so on. Well-designed Decision Management Systems keep track of deci- sions taken and outcomes achieved, then have the ability to make or rec- ommend automatic mid-course corrections to improve outcomes over

xv xvi Decision Management Systems time. Decision Management Systems provide competitive differentiation through every critical business processes, at each decision point, leading to optimized outcomes. I’m convinced that Decision Management Systems have the ability to deliver significant competitive advantage to businesses, governments and institutions. James does a thorough job of explaining the business value and the design elements of Decision Management Systems that are the enablers of a formidable business transformation. Deepak Advani Vice President, Business Analytics Products & SPSS, IBM Foreword by Pierre Haren

In the past 30 years, the evolution of computer science can be described as a constant effort to “reify,” a long march to transform all activities into “digital things.” We started with the structuring of data and the advent of relational database systems, which led to the ascension of Oracle; then with the reification of processes, with the Enterprise Resource Planning software wave leading to the emergence of SAP, and later of I2 for Supply Chain Management and Siebel Systems for Customer Relationship Management. We moved on to the Business Process Management wave, which now enables the description of most service activities into well-defined sequences of processes weaving human-based processes with computer- based processes. This BPM emergence sets the stage for the next reifica- tion wave: that of decisions. And this is what this book by James Taylor is about: how we can transform the fleeting process of decisions into digital things that we can describe, store, evaluate, compare, automate, and modify at the speed required by modern business. The rate of change of everything is the global variable, that has changed most over the last 30 years. Relational databases postulated the value of slow-changing table structures. Enterprise Resource Planning systems embedded best-of-breed processes into rather inflexible software architectures. However, nowadays, most decisions live in a very fast- changing environment due to new regulations, frequent catastrophic events, business model changes, and intensely competitive landscapes. This book describes how these decisions can be extracted, represented, and manipulated automatically in an AAA-rated environment: Agile, Analytic, and Adaptive.

xvii xviii Decision Management Systems

The long successful industrial experience of the author and his sup- porting contributors, and the diversity of their background, has enabled them to merge the points of view of business rules experts with predic- tive analytics specialists and operations research practitioners. This vari- ety of expert opinions on decisions and their reification has produced a very rich book sprinkled with real-life examples as well as battle-tested advice on how to define, implement, deploy, measure, and improve Decision Management Systems, and how to integrate them in the human fabric of any organization. The next area in the continuous integration of humans and computers in our modern world will be decisions. All decision-making managers— that is, every manager—should use this book to get ahead of the compe- tition and better serve their customers. Pierre Haren VP ILOG, IBM Preface

Decision Management Systems are my business and one of my passions. I have spent most of the last decade working on them. Four years ago I wrote Smart (Enough) Systems with Neil Raden, in which we laid the groundwork for talking about Decision Management Systems. I have spent the time since then working with clients and technology vendors to refine the approach. I have read a lot of books on business rules, data mining, predictive analytics, and other technologies. I have had a chance to work with lots of great people with deep knowledge about the technologies involved. And I have been for- tunate to work with many clients as they build and use Decision Management Systems. This book is the result. The book is aimed at those at the intersection of business and technology: executives who take an interest in technology and who use it to drive innova- tion and better business results, and technologists who want to use technol- ogy to transform the business of their organization. You may work for a company that has already built a Decision Management System, perhaps even many of them. More likely you work for an organization that has yet to do so. This book will show you how to build Decision Management Systems, give you tips and best practices from those who have gone before, and help you make the case for these powerful systems. I wrote the book the way I talk to my clients, trying to put on the page what I say and do when I am working with them. As a result, the book fol- lows the same path that most organizations do. It begins by setting a context and showing what is possible. By showing what others have done and discussing the Decision Management Systems that other organizations have built, the book draws out what is different about Decision Management Systems. By establishing that these systems are

xix xx Decision Management Systems agile, analytic, and adaptive, it shows how these differences allow Decision Management Systems to be used to transform organizations in critical ways. The core of the book describes the principles that guide the development of Decision Management Systems and lays out a proven framework for build- ing them. It shows you how to find suitable decisions and develop the under- standing of those decisions that will let you automate them effectively. It walks through how to use business rules, predictive analytics and optimiza- tion technology to build service-oriented components to automate these decisions. And it explains why monitoring and continuous improvement are so important to Decision Management Systems, and describes the processes and technology you need to ensure your Decision Management Systems per- form for the long haul. The book concludes with a set of people, process, and technology enablers that can help you succeed. The end result is a book that gives you the practi- cal advice you need to build different kind of information systems—Decision Management Systems.

James Taylor Palo Alto, California [email protected] Preface xxi

Approach

The objective of this book is to give the reader practical advice on why and how to develop Decision Management Systems. These systems are agile, ana- lytic, and adaptive—and they fundamentally change the way organizations operate. The book does not get into the details of every stage—it would have to be many times its length to do so—but focuses instead on the critical, practical issues of these systems. If you are not sure about the value proposition of Decision Management Systems, or have never come across them before, read Part I—Chapters 1-4. These chapters will introduce Decision Management Systems, and give you a sense of their importance to your organization. If you are already sure that you want to build Decision Management Systems, skip straight to Chapter 5 and read Part II—the core “how-to” part of the book. Don’t forget that first part, though—you will want to use it when building your business case! If you are about to embark on building a Decision Management System, check out the people, process, and technology enablers in Part III, Chapters 8-10, if you haven’t already.

How This Book Is Organized

This book is organized into three parts.

■ Part I: The Case for Decision Management Systems The first four chapters make the case for Decision Management Systems—why they are different and how they can transform a 21st cen- tury organization. ■ Chapter 1, “Decision Management Systems Are Different”: This chapter uses real examples of Decision Management Systems to show how they are agile, analytic, and adaptive. ■ Chapter 2, “Your business is your systems”: This chapter tackles the question of manual decision-making, showing how modern organizations cannot be better than their systems. ■ Chapter 3, “Decision Management Systems Transform Businesses”: This chapter shows that Decision Management Systems are not just different from traditional systems – they represent opportunities for true business transformation. xxii Decision Management Systems

■ Chapter 4, “Principles of Decision Management Systems”: By now you should understand the power of Decision Management Systems. This chapter outlines the key guiding principles for building them. ■ Part II: Building Decision Management Systems Chapters 5 through 7 are the meat of the book, outlining how to develop and sustain Decision Management Systems in your organization. ■ Chapter 5, “Discover and Model Decisions”: This chapter shows how to describe, understand, and model the critical repeatable decisions that will be at the heart of the Decision Management Systems you need. ■ Chapter 6, “Design and Implement Decision Services”: This chapter focuses on using the core technologies of business rules, predictive analytics, and optimization to build service-oriented decision-making components. ■ Chapter 7, “Monitor and Improve Decisions”: This chapter wraps up the how-to chapters, focusing on how to ensure that your Decision Management Systems learn and continuously improve. ■ Part III: Enablers for Decision Management Systems The final part collects people, process, and technology enablers that can help you be successful. ■ Chapter 8, “People Enablers”: This chapter outlines some key people enablers for building Decision Management Systems. ■ Chapter 9, “Process Enablers”: This chapter continues with process-centric enablers, ways to change your approach that will help you succeed. ■ Chapter 10, “Technology Enablers”: This chapter wraps up the enablers with descriptions of the core technologies you need to employ to build Decision Management Systems. ■ Epilogue ■ Bibliography Acknowledgments

First and foremost I would like to acknowledge the support of IBM. Deepak Advani and Pierre Haren were enthusiastic supporters of the book as soon as I proposed it. Mychelle Mollot, Brian Safron, and Erick Brethenoux helped close the deal with IBM Press and get the whole process kicked off. Many others were incredibly helpful during the pro- duction of this book. In particular, two IBM employees helped throughout the process. They supported me through the process, shared their thoughts and sug- gestions, helped me find other IBM experts in a number of areas, and made extensive direct contributions: Erick Brethenoux—Executive Program Director, Predictive Analytics & Decision Management Strategy, IBM. Erick’s responsibilities within IBM include mergers and acquisitions, strategic planning, predictive analytics corporate messaging, and future scenarios analysis. He also plays a major role in the industry analyst activities and various operational missions within the company. Erick was a VP of Corporate Development at SPSS, the predictive analytics company that IBM acquired in 2009. Prior to SPSS, Erick was VP of Software Equity Research at Lazard Frères, New York, and Research Director of Advanced Technologies at the Gartner Group. Erick has published extensively in the domains of artificial intelligence systems, system sciences, applied mathematics, complex systems, and cybernet- ics. He has held various academic positions at the University of Delaware and the Polytechnic School of Africa in Gabon. Jean Pommier—Distinguished Engineer & CTO, IBM Jean is a Distinguished Engineer and CTO in the IBM WebSphere Services organization and is in charge of Service Engineering (implemen- tation methods, best practices, and consulting offerings). Prior to xxiii xxiv Decision Management Systems

joining IBM in 2008, he was ILOG’s VP of Methodology. Jean joined ILOG upon its creation in 1987 in R&D, moving into consulting and then management in 1990. From 2003 to 2006, Jean led Worldwide Professional Services for ILOG; prior to that he headed worldwide con- sulting and U.S. sales operations for ILOG’s largest division. Jean has contributed to more than 400 successful customer implementations of Decision Management Systems. In addition, a number of IBM employees put their expertise to work helping me with specific sections. Many of them had to respond incredi- bly quickly so I could meet publishing deadlines and I could not have gotten the book done on time without them:

■ Jerome Boyer—Senior Technical Staff Member, BPM & BRM Implementation & Methodology ■ Dr. Asit Dan—SOA/BPM and Information Agenda Alignment, Chief Architect ■ Sarah Dunworth—Advisory Product Manager ■ Michael McRoberts—SPSS Chief Architect ■ Dr. Greger Ottoson—Product Management, Business Process & Decision Management ■ Vijay Pandiarajan—Product Marketing Manager, WebSphere Decision Management ■ David Pugh—Senior Manager, Product Management, Business Analytics (SPSS) ■ Bruno Trimouille—Program Director, WebSphere Worldwide Industry Marketing

In addition, a number of IBM staff helped me find suitable case study material:

■ Dr. Jeremy Bloom—Senior Product Marketing Manager, ILOG Optimization ■ Martha Mesa—Client Reference Manager, ILOG and WebSphere ■ Caroline Poser—WebSphere Client Reference Manager ■ Cheryl Wilson—BRMS Demand Generation Program Planning and Management Acknowledgments xxv

Pearson’s team was superb as always. Mary Beth Ray, Chris Cleveland, Kimberley Debus, and Jovana Shirley all excelled and made what was a compressed production schedule look easy. Thanks also to Steve Stansel for managing the IBM Press end of the process. I would also like to acknowledge the work of Dr. Alan Fish in the United Kingdom on decision dependency diagrams. Alan was generous with his time and ideas, and I for one am looking forward to his forth- coming book. Thanks to you all. Without you the book would be thinner, less accu- rate, and less complete. Any remaining mistakes are my own. This page intentionally left blank About the Author

James Taylor is the CEO of Decision Management Solutions, and is the leading expert in how to use business rules and analytic technology to build Decision Management Systems. James is passionate about using Decision Management Systems to help companies improve decision-making and develop an agile, analytic, and adaptive business. He has more than 20 years working with clients in all sectors to identify their highest-value opportuni- ties for advanced analytics, enabling them to reduce fraud, continually man- age and assess risk, and maximize customer value with increased flexibility and speed. In addition to strategy consulting, James has been a keynote speaker at many events for executive audiences, including ComputerWorld’s BI & Analytics Perspectives, Gartner Business Process Management Summit, Information Management Europe, Business Intelligence South Africa, The Business Rules Forum, Predictive Analytics World, IBM’s Business Analytics Forum, and IBM’s CIO Leadership Exchange. James is also a faculty member of the International Institute for Analytics. In 2007, James wrote Smart (Enough) Systems: How to Deliver Competitive Advantage by Automating Hidden Decisions (Prentice Hall) with Neil Raden, and has contributed chapters on Decision Management to multiple books, including Applying Real-World BPM in an SAP Environment, The Decision Model, The Business Rules Revolution: Doing Business The Right Way, and Business Intelligence Implementation: Issues and Perspectives. He blogs on Decision Management at www.jtonedm.com and has written dozens of articles on Decision Management Systems for CRM Magazine, Information Management, Teradata Magazine, The BPM Institute, BeyeNetwork, InformationWeek, and TDWI’s BI Journal.

xxvii xxviii Decision Management Systems

He was previously a Vice President at Fair Isaac Corporation, spent time at a Silicon Valley startup, worked on PeopleSoft’s R&D team, and as a consult- ant with Ernst and Young. He has spent the last 20 years developing approaches, tools, and platforms that others can use to build more effective information systems. He lives in Palo Alto, California with his family. When he is not writing about, speaking on or developing Decision Management Systems, he plays board games, acts as a trustee for a local school, and reads military history or science fiction. 4

Principles of Decision Management Systems

Organizations that have adopted Decision Management Systems have gained tremendous results from doing so. The use of business in Decision Management Systems has given organizations the agility to respond rapidly to competitive and market changes, to avoid business risks, and to take advantage of narrow win- dows of opportunity. The use of analytics to predict risk, fraud, and opportunity in these Decision Management Systems has kept companies profitable despite the risks they face and has allowed them to maximize the value of their customer relationships through a laser focus on opportunity. The ability of Decision Management systems to adapt to change and to be part of a learning environment has allowed organizations to experiment with new approaches, learn from their successes and failures, and continuously improve their business. Any organization would want the benefits of these kinds of systems. However, it is not immediately clear how to build Decision Management Systems. Although there are specific technologies involved, the use of these tech- nologies is not sufficient to ensure that Decision Management Systems are the outcome of using them. Decision Management Systems appear to deal with differ- ent issues, and have different characteristics, across different industries and business functions. It can be hard to see what an underwriting system in use by

47 48 Decision Management Systems the agents of a property and casualty insurer has in common with a real-time offer management system supporting a website. Yet when the basic principles of Decision Management Systems are understood, they can be correctly identified and delivered with maximum return on investment. Four specific principles are at the heart of identifying and building Decision Management Systems. If a system exists, it can be assessed against these four principles and can be said to be a Decision Management System if these principles guide its design and implementation. If an information systems project is being considered, then the integration of these principles into the project will ensure that what is delivered is a Decision Management System. The four principles address the characteristic capabilities of a Decision Management System:

1. Begin with the decision in mind. Decision Management Systems are built around a central and ongoing focus on automating decisions, particularly operational and “micro” decisions. 2. Be transparent and agile. The way Decision Management Systems make each decision is both explica- ble to non-technical professionals and easy to change. 3. Be predictive, not reactive. Decision Management Systems use the data an organization has collected or can access to improve the way decisions are being made by predicting the likely outcome of a decision and of doing nothing. 4. Test, learn, and continually improve. The decision-making in Decision Management Systems is dynamic and change is to be expected. The way a decision is made must be continually challenged and re-assessed so that it can learn what works and adapt to work better.

Principle #1: Begin with the Decision in Mind

Most information systems have been developed, and are continuing to be developed, around business functions, business data, or business processes. Functional systems support a set of related business functions such as accounting or human resources. Data-centric systems focus on 4: Principles of Decision Management Systems 49 particular kinds of data, such as customer information. Business processes such as order-to-cash have been layered on top of both kinds of systems, and newer systems are developed to deliver additional specific business processes. Each of these approaches has pros and cons, but they all share a common challenge—they assume either that people will make all the decisions involved in the functions and business processes being automated or that how these decisions are made can be fixed. As a result, none of them are Decision Management Systems. To develop Decision Management Systems, we must take a different approach—not one based on functions, data, or processes. To develop Decision Management Systems we must begin with the decision in mind. Decision Management Systems are built to automate and improve specific business decisions. As a decision involves making a selection from a range of alternatives, these systems make that selection—they choose the action or actions that can or should be made given the data available and the context of the decision. Decision Management Systems do not assume that every decision must always be taken by a human. Decision Management Systems make these decisions using the same business logic humans would apply without human intervention. Clearly, however, we are neither willing nor able to build information systems to make every decision on our behalf. Only certain decisions can and should be addressed by Decision Management Systems.

What Kinds of Decisions Are We Talking About? Information systems are good at handling repetitive tasks. They excel at doing the same thing over and over without variation and without making mistakes from one transaction to the next. Something that can- not be defined in a repeatable way is not a good target for any kind of information system; thus, only those decisions that are repeatable are good candidates for being automated and managed using a Decision Management System.

Repeatable Decisions A repeatable decision is one that is made more than once by an organ- ization following a well-defined, or at least definable, decision-making approach. Business decisions can be categorized in various ways; one effective way to look at decisions is to categorize them as strategic, tacti- cal, or operational (Taylor & Raden, 2007). This divides decisions into 50 Decision Management Systems three categories based on the value of each decision made—the differ- ence between a good and a bad decision—and the number of times such a decision is made by an organization:

■ Strategic decisions are those high-value, low-volume decisions that guide the overall direction of the company. These ad-hoc, typically one- off decisions are made by senior management or the executive team of an organization. Lots of information is assembled and analyzed while many options are considered. Once the decision is made, it is never made again in the same context—even if it is revisited later, this is really a different decision as circumstances are different. Organizations may know that a strategic decision is going to be needed well in advance, but often these decisions arise from unexpected opportunities or challenges. Strategic decisions are not candidates for Decision Management Systems as they lack the key element of repeatability. ■ Tactical decisions are those focused on management and control. These medium-value decisions still have significant business impact. They too involve data and analysis, typically by humans in management or knowl- edge worker positions. However, these decisions do repeat—the same kind of decision is made repeatedly during normal business operations. Decisions about the discounting approach being used or the staffing lev- els of a call center might be examples, and these decisions must be made every month or every week. The same or very similar analysis is per- formed each time, and company policies may play a significant role in how the decision is made. More repeatable and consistent tactical deci- sions are certainly targets for Decision Management Systems. ■ Operational decisions are those of lower individual value and typically relate to a single customer or a single transaction. They are critical to the effective operation of an organization, especially an organization of any size. Because of the number of times they must be made, consistency and repeatability are critical. Policies and well-defined decision making crite- ria are typically developed to ensure this consistency. Despite their low individual value, they are extremely valuable in aggregate. A decision made thousands or millions or even billions of times a year has a total value that often exceeds even the most important strategic decision. Furthermore, strategic and tactical decisions (for example, to focus on customer retention or discount more aggressively) will only have an impact if a whole series of operational decisions (how to retain this cus- tomer or what discount to offer this distributor) are made in accordance 4: Principles of Decision Management Systems 51

with the higher-level decision. For these reasons, operational decisions are the most common subject of Decision Management Systems.

To begin with the decision in mind, we must understand what opera- tional or tactical decision is to be the focus of our Decision Management System.

BUSINESS STRATEGY AND STRATEGIC DECISIONS This focus on repeatable operational and tactical decisions can and should be combined with a focus on business strategy. A business strategy must be supported by many operational and tactical decisions if it is to be put into practice. For instance, if a focus on growing per-customer revenue is central to your business strategy, then there may be strategic decisions that must be made to support this strategy. There will definitely be many operational and tactical decisions that will be influenced by and contribute to this strategy. For example, unless operational decisions about customer retention and cross-sell offers are made effectively, you cannot deliver on this customer-centric strategy. As discussed in Chapter 5, “Discover and Model Decisions,” the right operational decisions to focus on are those that support the objectives and key metrics of the organization.

Operational Decisions Operational decisions are by far the most common kind of repeatable decision. Every order placed, every customer interaction, every claim, or credit card transaction involves operational decisions. Operational deci- sions are the day-to-day, run-the-business decisions that are taken in large numbers by every organization. Operational decisions are highly repeatable—in fact, being consistent by following a set of guidelines or applying the relevant policies and regulations is a defining characteristic of an operational decision. Operational decisions can also involve an assessment of risk, as many forms of risk (loan default or credit risk, for instance) are acquired one transaction at a time. Operational decisions often must be made in real time or near real time, while customers are waiting for the decision to be made. Although many operational decisions are made about customers, they can also be made about shipments, suppliers, or staff. As more physical 52 Decision Management Systems devices are connected to the Internet with sensors or RFID chips, opera- tional decisions are often made about “things”—about vehicles, pack- ages, railcars, or network components.

Micro Decisions Micro decisions are a particular kind of operational decision (Taylor & Raden, 2007) where the desire to personalize an interaction with a cus- tomer requires a focus on making a decision for that customer and that customer only. Often an operational decision is repeated for all cus- tomers, with the decision being based only on the data available for the particular interaction or transaction concerned. A micro decision, in contrast, uses everything known or predictable about a customer to make a unique decision just for them. Two customers making the same request or involved in identical transactions would get two different outcomes. This focus on the information about the customer is what makes micro decisions a distinct form of operational decision. Everything known about the customer must be synthesized into actionable insight about the customer and fed into the operational decision alongside the information about the transaction. For example, when an order is placed, two operational decisions might be made—what shipping options to offer on the order and what discount to offer. The first of these might be managed as a standard operational decision with information about the order such as delivery address, weight, and value used to determine which of the various shipping options would be allowed. The second could be managed similarly but could also be handled as a micro deci- sion. The customer’s history with the company could be used to compute their likely future profitability and the risk that they might consider a competitor. This information, as well as information about the specific order, would then feed a micro decision to calculate a discount specific to this customer placing this order at this moment.

Tactical Decisions Although most repeatable decisions are operational or micro deci- sions, some are more tactical in nature. These repeatable tactical decisions often relate to management control of operations, such as assessing the staffing level required by a call center for the coming shift. They may also include knowledge worker decisions, such as those in clinical situations where a doctor is advised as to the likely interactions 4: Principles of Decision Management Systems 53 of a set of medications she has just prescribed to a patient. Although these tactical decisions are not as high-volume as operational or micro decisions, they are often of slightly greater value and so still offer an opportunity for Decision Management Systems.

CHANGING DECISION CRITERIA Another set of tactical decisions under managerial control are those for revising the decision criteria used in operational decisions. The world is dynamic, so the decision criteria for operational decisions need to change regularly. For example, customer preferences and fraud patterns change over time, as do the criteria for deciding what offers are to be made. Some tactical decisions are about setting the right decision criteria to be applied in an operational decision.

Because these systems are more complex and less repeatable, however, it is likely that the system will not completely automate the decision. Instead it will guide and support the decision maker by restricting the available options or by focusing them on a specific set of information, which will be useful to making the final determination.

Different Types of Decisions Interact Operational decisions are made every time a business process or trans- action executes, and tactical decisions are made periodically to change operational decision criteria or to attend to exceptional business situa- tions and take corrective actions. Over a longer horizon, this is not suffi- cient to improve business outcomes. Business strategy guides tactical and operational decisions and may need to change to respond to the dynamic marketplace and the external world. For example, competitors may introduce new products and services, influencing customer choice and putting competitive pressure on revenue and profitability. This may mean changing business strategy around providing targeted discounts and customizing products. This in turn creates change in processes and associated operational decisions. These strategic, tactical, and opera- tional decisions must be aligned. One way to understand the relationship between operational, tactical, and strategic decision-making is shown in Figure 4-1. This Observe- Orient-Decide-Act (OODA) model was originally introduced by military strategist and USAF Colonel John Boyd. Business outcomes are “observed” 54 Decision Management Systems

to detect changing situations that may lead to new tactical decisions, repre- sented by “decide” or strategy “reorientation” with changes in business processes and additional decisions. “Act” represents operational decisions following the decision criteria set by the “decide” stage. There’s more on the OODA Loop in Chapter 10, “Technology Enablers.”

Run the business Decide Act Know and choose Make operational Decision what to do decisions based on Management actual situation Systems Determine and Improve Execute and Automate

Orient Observe Make sense of Watch what is the big picture happening/has happened

Define strategy Assess and Monitor and Decision improvement Analyze Report (offline)

Based on IDC's Decision Continuum Model and the OODA Military Decision Loop

Figure 4-1 The Decision Lifecycle From strategy definition to decision automation.

If We Are Talking About Decisions, Aren’t We Just Talking About Decision Support Systems? The line between Decision Management Systems and Decision Support Systems (or Executive Information Systems) can be blurry. This is especially true when considering the kind of Decision Management System that handles tactical decisions, or where an operational decision is not completely automated—where the user is presented with multiple valid options, such as possible offers to make. 4: Principles of Decision Management Systems 55

Decision Management Systems are distinct, however, and they differ from traditional Decision Support Systems in five ways:

1. Decision Support Systems provide information that describes the situation and perhaps historical trends so that humans can decide what to do and which actions to take. Decision Management Systems automate or recommend the actions that should be taken based on the information that is available at the time the decision is being made. 2. The policies, regulations, and best practices that determine the best action are embedded, at least in part, in a Decision Management System where a Decision Support System requires the user to remember them or look them up separately. 3. The information and insight presented in a Decision Support System is typically backward looking, and Decision Support Systems are generally reactive—helping human decision-makers react to a new or changed situation by presenting information that might help them make a decision. In contrast, Decision Management Systems use information to make predictions and aim to be proactive. 4. Learning is something that happens outside a Decision Support System and inside a Decision Management System. Users of Decision Support Systems are expected to learn what works and what does not work and to apply what they learn to future decisions. Decision Management Systems have experimentation or test-and- learn infrastructure built in so that the system itself learns what works and what does not. 5. Decision Management Systems are integrated into an organization’s runtime environment. They make decisions for applications and services in the organization’s enterprise application architecture. In contrast, Decision Support Systems are often desktop or interactive applications that execute outside the core application portfolio.

Why Don’t the Other Approaches Work? Before considering the remaining principles, it is worth considering why it is essential to begin with the decision in mind. What is it about a focus on functions, on process, or on data that prevents the effective development of Decision Management Systems? 56 Decision Management Systems

A Functional Focus Is Not Enough One traditional approach to building systems is to focus on a cluster of related functions—those to do with human resources or those to do with managing a factory, for instance. Such systems contain stacks of capability focused in one functional area and owned by a single func- tional department. This approach could result in the development of Decision Management Systems if the decisions involved were wholly contained within a single business function. However, while some deci- sions are concentrated in this way, many cut across functions. A discount calculation decision, for instance, might involve inputs from supply chain functions, from finance, and from customer management. As such a focus on functions will rarely identify and encompass decisions in a way that lends itself to the construction of Decision Management Systems.

A Process Focus Is Not Enough Functional applications have gradually fallen from favor as organiza- tions have moved to focus on end-to-end business processes. Business processes such as “order to cash” or “issue policy” often cut across several functional areas, linking elements of one function together with ele- ments of another to create a useful business outcome. Although this cross-functional approach can help with the identification of decisions, a pure process focus tends to entwine decisions with the process itself. If no real distinction is drawn between decisions and the processes that need those decisions, it is hard to create true Decision Management Systems. A strong separation of concerns, keeping business processes and decisions linked but separate, is required if enough of a focus on deci- sions is to be maintained. Some processes keep decisions separate and manage them separately by assigning these decisions to people in manual process tasks. A focus on human decision-making, even in high-volume operational processes, also does not result in the construction of Decision Management Systems.

A Data Focus Is Not Enough Particularly when constructing their own custom systems, organiza- tions often focus on the data that must be managed. These systems become focused almost entirely on the management of the data elements or entities concerned. Providing what is known as CRUD functionality 4: Principles of Decision Management Systems 57

(Create, Read, Update, and Delete) for the core objects becomes their rationale. The data contained is managed only so that it can be edited and displayed while analysis is limited to reporting. Such systems often provide data for decision support systems but they, like process- and function-centric systems, defer decision making to actors outside the system.

Principle #2: Be Transparent and Agile

Most information systems in use today are opaque and hard to change. The use of programming languages—code—to specify their behavior makes them opaque to any but the most technically adept. This opacity, and the difficulties of confirming that changes made to the code do what they are expected to do, make for long change cycles and a lack of responsiveness. The combination means that extensive information tech- nology projects must be planned, budgeted, and executed to make changes to the behavior of a system. These characteristics are unacceptable in a Decision Management System. Opacity is unacceptable because many decisions must demon- strate that they are compliant with policies or regulations. If the code is opaque, then it will not be possible to see how decisions have been made and it will not be possible to verify that these decisions were compliant. Decision Management Systems also make decisions that are based on detailed business know-how and experience. If the code is so opaque that it cannot be understood by those who have this know-how or experience, then it is unlikely to be correct. Organizational decision-making changes constantly, so agility is also essential. As regulations change, the behavior of any Decision Management System that implements that regulation must also change. Organizations also want Decision Management Systems to make good decisions—effective ones. Effective decisions based on the expectations of customers must be competitive, yet the behavior of competitors and customer expectations change constantly. And moreover, customers and competitors are not obliged to tell organizations when their expectations or plans change. An ability to rapidly change Decision Management Systems to respond is essential. Decision Management Systems must therefore be both transparent and agile: 58 Decision Management Systems

■ The design must be transparent so that it is clear that the system is exe- cuting the behavior expected of it. ■ The execution must be transparent so that it is clear how each decision was made. ■ They must be agile so that their behavior can be changed when necessary without delay and without unnecessary expense.

Design Transparency and Why It Matters A Decision Management System must exhibit design transparency. It must be possible for non-technical experts—those who understand the regulations or policies involved or who have the necessary know-how and experience—to determine whether the system is going to behave as required. Those without IT expertise must be able to manage the way in which decisions are made so that it is clear to all participants involved. The drivers or source of this behavior must be identifiable so that those reviewing the behavior of the system can clearly assess its effectiveness in meeting objectives. Tracking the source of decision-making behavior also means that changes in those sources can be quickly mapped to the changes required in the system. Design transparency means it is possible to determine the way in which a proposed change will ripple through the Decision Management System. One regulatory change might affect many deci- sions, for example, and decisions may be dependent on the same data elements because they have information needs in common. Organizations must be sure that their Decision Management Systems will make decisions accurately and effectively after a change is made. This requires that the ripples and impacts of any change can be deter- mined before it is made. Design transparency is essential to being able to trace these impacts.

Execution Transparency and Why It Matters When a decision is made by a person, that person can be asked to explain the decision. If a person rejects an application for a loan, for instance, he can be asked to appear in court, to write a letter explaining, or simply to answer the customer’s questions. This is not possible when 4: Principles of Decision Management Systems 59 a system makes a decision. A Decision Management System must there- fore provide an explanation of a decision that will satisfy customers or suppliers who are materially affected by it. When a decision is regu- lated, such as when deciding which consumers may have access to credit, a Decision Management System must provide an exact description of how each decision was made so that it can be reviewed for compliance. Decision Management Systems must deliver real execution transparency in these cases. Not all decisions require execution transparency. When marketing or promotional decisions are being made, it may not be necessary to under- stand exactly why a particular offer was made to a particular site visitor. When a Decision Management System is being used to decide when to bring a human into the loop, for fraud investigation for instance, it may also not be necessary to understand why as the human acts as a second “pair of eyes.” Even when a Decision Management System does not require execution transparency, an understanding of how each decision was made can help improve the decision-making of the system. Building in execution transparency is therefore generally a good idea, whether it is required or not. Any approach to developing Decision Management Systems must support execution transparency as well as design transparency.

Business Agility and Why It Matters An increase in transparency is likely to result in an increase in busi- ness agility—if it is easier to see how something works, it will be easier to change how it works when this is needed. A faster response to a needed change improves overall business agility. Transparency is neces- sary for agility but not sufficient. Once a change is identified and its design impact assessed, it must be possible to make the change quickly and reliably. Decision Management Systems can require real-time changes to their behavior in extreme cases. Daily or weekly changes are very common. When sudden market changes occur, such as major bank- ruptcies or an outbreak of hostilities, the resulting need for changes to Decision Management Systems can be extreme. Money—and perhaps lives—will be lost every minute until the change is made. Decision Management Systems must change constantly to reflect new regulations, new policies, and new conditions. This rate of change must 60 Decision Management Systems

be both possible and cost-effective. For most businesses and other organ- izations, it will not be acceptable if the needed agility in Decision Management Systems comes at too high a price. For a Decision Management System, change must be easy, it must be reliable, it must be fast, and it must be cost-effective.

AGILE DECISION-MAKING FOR TRULY AGILE PROCESSES Many organizations invest a great deal in developing agile business processes. Decision Management Systems further increase this agility as business changes often involve updates to business decisions. These decisions are often the most dynamic part of a process, the part that changes most often. For instance, a company’s pricing rules are likely to change far more often than its order-to-cash process. If only the business process can be changed quickly, then the company will not be able to respond to the far more numerous pricing changes without changing its process, an unnecessary step. Developing Decision Management Systems allows an organization to control business processes and the critical decisions within them. This increases the agility built into a process and allows for a stable process even when decision-making is constantly changing and evolving. Explicitly identifying decisions and describing the logic behind them allows this logic to be managed and updated separately from the process itself, dramatically increasing the agility of an organization.

Principle #3: Be Predictive, Not Reactive

In recent years, organizations have spent heavily on technology for managing and using data. Beginning with Database Management Systems and moving through Information Management, Data Quality and Data Integration to Reporting, and Dashboards, these investments are now mostly classified as Business Intelligence and Performance Management. These investments have taken data that was once hidden in transactional systems and made it accessible and usable by people making decisions in the organization. These investments have been focused on analyzing the past and pre- senting this analysis to human users. They have relied, reasonably enough, on their human users to make extrapolations about the future. Users of these systems are making decisions based on this data, using 4: Principles of Decision Management Systems 61 what has happened in the past to guide how they will act in the future. Many of these systems can also bring users’ attention to changes in data quickly to prompt decision-making. The value of this investment in terms of improved human decision-making is clear. These approaches will not work for Decision Management Systems. When a decision is being automated in a Decision Management System, there is no human to do the extrapolation. Passing only historical data into a Decision Management System would be like driving with only the rear view mirror—every decision being made would be based on out-of- date and backward-looking data. It fact it would be worse, as a human driver can make guesses as to what’s in front of her based on what she sees in a rear view mirror. She will be reasonably accurate too, unless the road is changing direction quickly. Systems are not that smart—without people to make extrapolations from data, Decision Management Systems need to be given those extrapolations explicitly. Without some view of the future and the likely impacts of different decision alternatives, a Decision Management System will fail to spot opportunities or threats in time to do anything about them. Predicting likely future behavior is at the core of using predictions in Decision Management Systems. You need to predict individual customer behavior such as how likely they are to default on a loan or respond to a particular offer. You need to predict if their behavior will be negative or positive in response to each possible action you could take, predicting how much additional revenue a customer might generate for each possi- ble action. You want to know how likely it is that a transaction repre- sents risky or fraudulent behavior. Ultimately you want to be able to predict the best possible action to take based on everything you know by considering the likely future behavior of a whole group of customers. Decision Management Systems require predictions. They must be given predictions in the context of which they can act instead of simply reacting to the data available at the time a decision is made. They need access to predictions that turn the inherent uncertainty about the future into a usable probability. They cannot be told, for instance, which claims are definitely fraudulent—this is uncertain. They can be given a model that predicts how likely it is that a specific claim is fraudulent. There are three specific ways in which Decision Management Systems can be given predictions. They can be given models that predict risk or fraud, that predict opportunity, and that predict the impact of decisions. They can use these predictions to direct, guide, or push decision-making in the right direction. 62 Decision Management Systems

Predict Risk or Fraud Most repeatable decisions do not have a huge economic impact indi- vidually. Despite their limited scope, many do have a significant gap between good and bad decisions. The value of the decision varies signifi- cantly with how well they are made. This gap arises when there is a risk of a real loss if a decision is made poorly. For instance, a well-judged loan offer to someone who will pay it back as agreed might net a bank a few tens of dollars in profit. A poorly judged offer will result in the loss of the loan principal—perhaps thousands of dollars. This mismatch between upside and downside is characteristic of risk-based decisions. Similarly, a poorly made decision in detecting fraud can result in large sums being transferred to an imposter or large purchases being made using stolen credit cards. When Decision Management Systems are being used to manage these decisions, it is essential that the decision-making be informed by an accurate assessment of the risks of the particular transaction or customer concerned. Such models might be focused on fraud, using analysis of patterns revealed in past fraudulent transactions to predict how likely it is that this transaction is also fraudulent. They might be focused on the likelihood of default, using a customer’s past payment history and the history of other customers like him to predict how likely it is that he will fail to make payments in a timely fashion. Many techniques can be used to build such models from historical data, but all of them require knowledge of which historical transactions were “bad”—fraudulent or in default. These known cases are used to train a model to predict how similar a new transaction is to these “bads.” Once such a prediction exists, a Decision Management System can use it, treating those transactions or customers with particularly high, or par- ticularly low, risk differently.

Predict Opportunity Many decisions do not involve an assessment of downside risk, but they still have some variability. Not driven entirely by compliance with regulations or policies, these decisions require an assessment of opportu- nity before an appropriate choice can be made. There is typically no absolute downside if a poor decision is made, simply a missed opportu- nity. When Decision Management Systems are being used to manage these opportunity-centric decisions, they will need to have some way to manage these tradeoffs. 4: Principles of Decision Management Systems 63

These decisions are largely, though not exclusively, about how to treat customers. Deciding which offer to make to a customer or which ad to display to a visitor are examples of decisions where the “best” decision is one which makes the most of the opportunity to interact with the cus- tomer or visitor. Historical data can be used to predict how appealing a particular offer or product might be to a particular person or to a spe- cific segment of customers. The value to the company of each offer, combined with the likelihood that a particular customer will accept it, can then be used to identify the most effective offer—to make the best decision. When many such offers are being considered, it may be complex to identify the “best” offer. It may be difficult to manage the tradeoffs between the various decisions. In these circumstances, Decision Management Systems can take advantage of optimization technology that allows the tradeoffs to be explicitly defined, and then the “optimal” or best outcome can be selected mathematically.

Predict Impact of Decisions Sometimes the effect of an action taken by a Decision Management System cannot be precisely determined. For instance, the value of a sub- scription for a mobile phone will vary with the use made of the phone. When an action is available for a decision and has this kind of uncer- tainty about its value, a further prediction is needed. The likely impact of each action on the profitability, risk or retention of a customer can be predicted by analyzing the behavior of other similar customers who were treated the same way—for whom the same action was taken. The prediction of the likely impact of each action can be combined with predictions of risk and opportunity to improve the qual- ity of decision-making in Decision Management Systems.

Principle #4: Test, Learn, and Continuously Improve

Most information systems have a single approach to handling any decisions that have been embedded in them. Every transaction is treated the same way, with possible alternative approaches largely eliminated during design to find the “best” approach. Once this singular approach has been implemented, information systems continue to work the way they were originally designed until someone explicitly re-codes them to behave differently. The only way these systems are changed is when an 64 Decision Management Systems external agent—a human—decides that a change is required. These sys- tems also accumulate large amounts of data about customers, products, and other aspects of the business. This data might show that certain actions are more effective than others, but the system will continue with its programmed behavior regardless—every customer is treated like the first. This approach is not an effective way to develop Decision Management Systems. When we make decisions about our own lives or interactions, we often assess a large amount of data, either explicitly or implicitly. We learn from this data what is likely to work or not work— the data accumulated provides clues to how an effective decision can be made. A Decision Management System cannot afford to ignore the accu- mulated historical data. Decisions involve making a selection from a range of alternative actions and then taking the selected action. It is often not immediately obvious if the decision was made effectively. Some decisions have a sig- nificant time to outcome, and no assessment of the effectiveness of the decision will be possible until that time has passed. For instance, an early intervention designed to ensure a customer renews her annual con- tract cannot be assessed until the customer reaches the renewal point, perhaps many months later. If the action taken turns out to be ineffec- tive, then a different approach will need to be considered. A Decision Management System cannot afford to “single thread” this analysis by only testing one decision making approach at a time. Whether a decision is a good one or a bad one is a moving target. A decision may be made to discount a particular order for a customer that may be competitive today but much less so tomorrow because a competi- tor has changed their pricing. As markets, competitors, and consumer behavior shift, they affect the effectiveness of a decision. This constant change in the definition of an effective decision means that Decision Management Systems must optimize their behavior over time, continu- ously refining and improving how they act. Decision Management Systems must therefore test, learn, and contin- uously improve. The analysis and changes may be done by human observers of the Decision Management System or by the system itself in a more automated fashion. Decision Management Systems must collect data about the effectiveness of decision making. They must use this data, and other data collected by traditional information systems, to refine and improve their decision-making approach. Decision Management 4: Principles of Decision Management Systems 65

Systems must allow multiple potential decision-making approaches to be tried simultaneously. These are continually compared to see which ones work and which ones do not. Successful ones persist and evolve, unsuccessful ones are jettisoned. Finally, Decision Management Systems must be built on the basis that their behavior will change and improve over time. Decision Management Systems will not be perfect when implemented but will optimize themselves as time passes.

Collect and Use Information to Improve The first way Decision Management Systems must learn is through collecting and then using information about the decisions they make. When a Decision Management System makes a decision, it should record what decision it made, as well as how and why it made the decision it did. This decision performance information will allow the long-term effectiveness of a decision to be assessed as it can be integrated with the organization’s performance metrics to see which decisions result in which positive, or negative, performance outcomes. This information allows good decisions to be differentiated from bad ones, better ones from worse ones. It is often said that if you wish to improve something, you must first measure it. Decisions are not an exception to this rule. Information about the decisions made can and should be combined with the information used to make the decision. This information might be about a customer, a product, a claim, or other transaction. This is the information that is passed to the Decision Management System so that it can make a decision. Combining this information with the decision per- formance information will identify differences in performance that are caused by differences in the information used to drive the decision. For instance, a decision-making approach may work well for customers with income below a certain level and poorly for those above it. Storing, inte- grating, analyzing, and using this data to improve decision-making is the first building block in building Decision Management Systems that continuously improve.

Support Experimentation (Test and Learn) When a Decision Management System is being defined, it may not be clear what approach will result in the best outcomes for the organiza- tion. Several alternative approaches might all be valid candidates for 66 Decision Management Systems

“best approach.” Simulation and modeling of these approaches, and test- ing them against historical data, might show which approach is most likely to be superior. Even if the historical data points to a clear winner, the approach is going to be used against new data and may not perform as well in these circumstances. A Decision Management System, therefore, needs to be able to run experiments, choosing between multiple defined approaches for real transactions. The approach used for each transaction can be recorded, and this information will allow the approaches to be compared to see which is superior. This comparison may not be definitive, and one approach may be better for some segments of a customer base, while a second works better for other segments. Results from these experiments can then be used to update the Decision Management System with the most successful approach or combination of approaches. Because Decision Management Systems handle repeatable decisions, there will always be more decisions to be made that will be able to take advantage of this improved approach.

Optimize Over Time In a static world, one round of experimentation might be enough to find the best approach. A set of experiments could be conducted and the most effective approach selected. As long as nothing changes, this approach will continue to be most effective. However, the effectiveness of a decision-making approach can vary over time for many reasons, and you have little or no control over this. The old “best” approach may degrade suddenly or gradually, and when it does, you will need to have alternatives. Even when experimentation finds a clear winner, a Decision Management System needs to keep experimenting to see whether any of the alternative approaches have begun to outperform the previous win- ner. Alternatives approaches could be those rejected as inferior initially or new ones developed specifically to see whether a new approach would be superior in the changing circumstances. The effect of this continuous and never-ending experimentation is to optimize results over time by continually refining and improving decision-making approaches. 4: Principles of Decision Management Systems 67

Summary

Decision Management Systems are different from traditional informa- tion systems.

■ Traditional information systems have a process, data, or functional focus. Decision Management Systems are decision-centric, built with a repeat- able decision in mind. ■ Traditional information systems are opaque and hard to change. Decision Management Systems improve collaboration and compliance by being transparent and agile. ■ Traditional information systems present historical data as analyses to peo- ple. Decision Management Systems embed analytics that predict risk, opportunity, and impact deep into the system itself. ■ Traditional information systems are static and don’t use the data they store to improve their results. Decision Management Systems test new approaches, learn what works, and continuously improve.

Developing Decision Management Systems requires a new approach; this is the subject of Part II, “Implementing Decision Management.” Index

Numbers Aeroports de Paris, 14 agility, 4 24/7 world, consumer business agility, 59-60 expectations, 23 changing circumstances, 5-6 compliance, 6 A process improvement, 7 A/B testing, 173 airline planning and operations, 17 ABRD (Agile Business Rule aligning around business results, Development), 215-216 three-legged stool, 195 act, OODA loop, 234 always on, 33-35 adapting SDLC (software banking, 34 development lifecycle), 215-219 insurance, 34-35 multiple data models and business analytic data marts, 247, 250 process management, 220 analytics, 8 adaptive, 15 focusing resources, 13-14 finding new approaches, 15-16 improving decision-making by using managing tradeoffs, 17-18 data, 15 testing and learning, 16-17 in-database analytics, 141 advanced simulation, impact of managing risk, 8-10 decision-making approaches, 187 reducing fraud, 10-11 adverse selection, 9 targeting and retaining, 11-13

273 274 Decision Management Systems analytics people, 192 required elements, 237-238 analytics teams, 192 templates, 134 analyzing build test and learn infrastructure, legacy systems, 95-96 121-122 past data, 97 building predictive analytic root cause, 98 models, 139 what-ifs, 98 business agility, 59-60 assessing business experimentation, 224 decision effectiveness, 170-172 business impact, 79 design impact, 168-169 business processes, decision assets, maximizing, 41-44 inventory, 212 association models, 240 business rules association rule algorithms, 127 building Decision Services, 123 automation, models, 241 data, 127-129 deploying and integrating B continuously, 137 executable rules, 130-131 Bachelard, Gaston, 31 experts, 126 Bancolombia, 10-11 legacy systems, 124-125 banking policies and regulations, 125-126 always on, 34 source rules, 123-124 breaking ratios, 37 data mining, 135 fraud, 40 traceability, 136 Benecard, 6 validating, 136 best practices verifying, 136 decision services Business Rules Management System designing for batch and interactive, (BRMS), 130-132, 218-219, 152-153 235-236 iterative development, 155 building rule management logging, 154-155 environments, 177 no side effects, 153-154 required elements, 237-238 decision-centric data, 250 templates, 134 Bharti Airtel, 41 business rules representation, Big Data, scale of business selecting, 132-134 operations, 24 big data platforms, 253-255 blocking, 227 C Boyd, Colonel John, 53 calculation, 82 BPMS (Business Process Management candidates for automation, System), 257-258 characteristics of suitable breaking ratios, 36-38 decisions, 80-81 BRMS (Business Rules Management Carrefour Group, 11-13 Systems), 130-132, 218-219, case management integration, 235-236 Decision Services, 151-152 building rule management Center of Excellence (CoE), 196-206 environments, 177 Index 275 champion-challenger, predictive real-time responsiveness, 20-21 analytic models, 180 self-service, 22-23 champion-challenger testing, 174 control groups, experimentation, characteristics of suitable 224-225 decisions, 72 credit cards, fraud, 40 candidates for automation, 80-81 cross-training, 194 measurable business impact, 79 customers, market of one, 31 non trivial, 75 experience of one, 32 analysis is required, 77 paralysis of choice, 31 continued updates, 78 domain knowledge and expertise, 76 D large amounts of data, 77 data large numbers of actions, 78 analyzing past data, 97 policies and regulation, 76 business rules, Decision Services, trade-offs, 78 127-129 repeatability, 72 capturing decision effectiveness consistent measurement of success, 75 data, 164 decision is made at defined times, 73 changes to, 161-162 defined set of actions, 74 decision service execution data, 164 same information is used each time, 73 decision service response data, 164 cleaning up data, 242 enterprise data, 165 Clinical Decision Support systems, 45 preparing for predictive analytic clustering models, 240 models, 138-139 code, table-driven code, 95 scale of big business, 24 collaboration data cleanup, 242 decision inventory, 213 data focus, 56 three-legged stool, 193-194 data infrastructure, 247 comparing analytic data marts, 250 alternative approaches, 182-183 big data platforms, 253-255 approaches, 172 data warehouses, 249 comparison, 226 in-database analytics, 251-253 compliance, agility, 6 operational databases, 247-248 confirming impact of decision- data integration, Decision Services, making approaches, 184 148-149 advanced simulation, 187 data mining, business rules, 127, 135 simulations, 185 data warehouses, 249 testing, 184 Davenport, Tom, 228 what-if analysis, 186 decide, OODA loop, 234 consistency, decision decision analysis, 158 characteristics, 109 decision characteristics, 108 constraints, optimization models, 144 consistency, 109 consumer expectations degrees of freedom, 110 24/7 world, 23 time to value, 110 global services, 22 276 Decision Management Systems

timeliness, 109 begin with a decision in mind, 48-49 value range, 109 Decision Support Systems, 54-55 volume, 108 micro decisions, 52 Decision Discovery, three-legged operational decisions, 51 stool, 194-195 repeatable decisions, 49-50 decision effectiveness, assessing, tactical decisions, 52-53 170-172 test, learn, and continuously improve, decision effectiveness data, 63-65 capturing, 164 collect and use information to improve, 65 decision inventory, managing, optimize over time, 66 211-212 support experimentation, 65 business processes, 212 decision monitoring environments, collaboration, 213 building, 165-166 linking to implementation, 214-215 decision points, process-centric, 89 refactoring on each project, 213 decision service execution data, 164 synchronizing with performance Decision Service integration patterns, management changes, 214 221-222 decision management, linking with decision service response data, 164 performance management, 167 Decision Services, 116 Decision Management Center of best practices Excellence, 196-206 designing for batch and interactive, Decision Management Systems, 3-4 152-153 overview, 263-265 iterative development, 155 pre-configured, 244-245 logging, 154-155 cons of, 246 no side effects, 153-154 pros of, 245 building business rules, 123 three-legged stool, 192-193 data, 127-129 aligning around business results, 195 deploying and integrating building collaboration skills, 193-194 continuously, 137 Decision Discovery, 194-195 executable rules, 130-131 resolving organizational issues, 196 experts, 126 starting three-way conversations, 193 legacy systems, 124-125 Decision Management Systems policies and regulations, 125-126 principles, 48 source rules, 123-124 be predictive not reactive, 60-61 building optimization models, 141-143 predicting impact of decisions, 63 constraints, 144 predicting opportunity, 62 decision variables, 144 predicting risk or fraud, 62 iterate, 147 be transparent and agile, 57 objectives, 144 business agility, 59-60 solvers, 146 design transparency, 58 building predictive analytic models, execution transparency, 58-59 137-138 building and testing, 139 deploying, 140 Index 277

exploring and understanding data, 138 linking to the business, 101 preparing data, 138-139 modeling dependencies, 102-108 selecting techniques, 139 eligibility, 81-82 building scaffolding, 116-117 finding, 87-88 build test and learn infrastructure, analyzing legacy systems, 95-96 121-122 event-centric, 91-94 defining service contracts and test cases, metrics and performance management 117-119 systems, 96-98 turning dependency networks into flow, process-centric, 89-91 119-121 top-down, 88 integration, 147 fraud decisions, 84-85 case management integration, 151-152 mapping, 96-97 data integration, 148-149 micro decisions, 52, 86-87 event integration, 150 monitoring, 159 process integration, 150 proactive changes, 163-167 performance, 171 reactive changes, 159-162 versioning, 256 operational decisions, 50-51 Decision Support Systems, 54-55 opportunity decisions, 85 decision-centric data, best prioritizing, 111-113 practices, 250 repeatable decisions, 49-50 decision-making repeatable management decisions, 86 improving, 8 risk decisions, 83 improving by using data, 15 strategic decisions, 50-51 decision-making approaches tactical decisions, 50, 52-53 A/B testing, 173 validation, 82 champion-challenger testing, 174 dependencies deploying changes, 187 determining, 102 developing new, 176 finding, 104 building rule management environments, modeling, 102-104 177-179 finding information sources, 104-105 comparing alternatives, 182-183 finding know-how, 106-108 designing new approaches, 181 iterate, 108 updating predictive analytic models, dependency loops, 104 179-180 dependency networks, turning into fact-based decisions, 228 flow, 119-121 refining with optimization, 174-176 deploying decision tables, 132 business rules, Decision Services, 137 decision trees, 133 changes to decision-making decision variables, optimization approaches, 187 models, 144 predictive analytic models, 140 decisions, 49 depressions, 208 calculation, 82 design documenting, 99 assessing impact, 168-169 decision basics, 99-100 experimentation, 226-227 decision characteristics, 108-110 design impact analysis, 168 278 Decision Management Systems design tools executable rules for modelers, 242 business rules, Decision Services, optimization systems, 244 130-131 design transparency, 58 objects, 131 designing new decision-making execution transparency, 58-59 approaches, 181 expectations, 20 developing decision-making 24/7 world, 23 approaches, 176 global services, 22 building rule management real-time responsiveness, 20-21 environments, 177-179 self-service, 22-23 comparing alternatives, 182-183 experimental designs, 182 designing new approaches, 181 experimentation, 222-223 updating predictive analytic models, business experimentation, 224 179-180 control groups, 224-225 distributed interactions, 27 design, 226-227 documenting decisions, 99 IT experimentation, 223-224 decision basics, 99-100 supporting, 65 decision characteristics, 108-110 experts linking to the business, 101 Decision Services, 126 modeling dependencies, 102-108 fact-based decisions, 230-232 external data costs, managing, 106 E external sources, 105 e-government initiatives, 33 effectiveness, assessing decision F effectiveness, 170-172 fact-based decisions efficiency, scale of big business, 24-25 decision-making approaches, 228 eligibility, 81-82 information strategies, 232 enriching live data, 260 moving to, 228 energy networks, 42 presenting data with decisions enterprise data, 165 together, 230 Equifax, 16 role of experts, 230-232 escalation, process-centric, 91 statistical awareness, 229-230 event-centric, finding decisions, 91 factorial experiments, 227 event and non-event data, 92-93 finding event streams, 93-94 decision dependencies, 104 finite state analysis, 94 decisions, 87-88 event data and non-event data, 92-93 analyzing legacy systems, 95-96 event integration, Decision event-centric, 91-94 Services, 150 metrics and performance management Event Processing System, 258-261 systems, 96-98 event streams, event-centric, 93-94 process-centric, 89-91 executable business rules, top-down, 88 writing, 134 information sources, 104-105 know-how, 106-108 finite state analysis, event-centric, 94 Index 279 flow, turning dependency networks testing, 184 into, 119-121 what-if analysis, 186 focus, 56 impact of decisions, predicting, 63 focusing resources, 13-14 improving decision-making, 8 fraud by using data, 15 banking, 40 in-database analytics, 141, 247, credit cards, 40 251-253 crushing, 39-41 incremental patterns, 222 insurance, 39 Infinity Insurance predicting, 62 breaking ratios, 36 reducing, 10-11 fraud, 40 fraud decisions, 84-85 information sources, finding, 104-105 freedom, decision characteristics, 110 information strategies, fact-based freeform rules, 133 decision, 232 functional focus, 56 informed pessimism, 208 insurance G always on, 34-35 fraud, 39 Getronics, 15 underwriters, 36 Gladwell, Malcolm, 230 integrating business rules, Decision global services, consumer Services, 137 expectations, 22 integration, Decision Services, 147 goals, 111 case management integration, 151-152 government, always on, 33 data integration, 148-149 graphs, 134 event integration, 150 Grup Bancolombia, 10-11 process integration, 150 interactions, 25 H distributed interactions, 27 HDFS (Hadoop Distributed file mobile interactions, 26 System), 253 social interactions, 26-27 healthcare, 46 interactive user interfaces, Clinical Decision Support systems, 45 optimization systems, 244 HealthNow New York, 7 Interactive Voice Response HIPAA (Health Insurance Portability systems, 24 and Accountability Act), 6 InterConnect (Equifax), 16 internal sources, 105 I-J IT departments, 192-193 IT experimentation, 223-224 IBM Microelectronics, 43 iterate impact analysis, 170 decision dependencies, 108 impact of decision-making, optimization models, 147 confirming, 184 advanced simulation, 187 simulations, 185 280 Decision Management Systems

K mathematical optimization, 243 maximizing know-how, finding, 106-108 assets, 41-44 KPIs, 94-96 revenue, 44-45 mapping, 96-97 measurable business impact, KPN, adaptive, 15-16 characteristics of suitable decisions, 79 L member enrollment processes, 7 learning, 16-17 metrics, 96 legacy systems, 95-96 analyzing past data, 97 Decision Services, 124-125 mapping decisions and KPIs, 96-97 linking root cause analysis, 98 performance management and decision micro decisions, 52, 86-87 management, 167 mobile devices, 26 to implementation, decision inventory, mobile interactions, 26 214-215 model refresh, predictive analytic to source rules, 135 models, 179 live data capture, Event Processing model repository, 241 System, 259 modelers, design tools, 242 local exceptions, process-centric, 90 modeling algorithms, 242 loops, dependency loops, 104 modeling dependencies, 102-104 Loveman, Gary, 225 finding information sources, 104-105 loyal customers, 27 finding know-how, 106-108 iterate, 108 M modeling languages, optimization systems, 244 managing models decision inventory, 211-212 association models, 240 business processes, 212 choosing which one to use, 241 collaboration, 213 clustering, 240 linking to implementation, 214-215 monitoring, 163 refactoring on each project, 213 predictive analytic models, 240 synchronizing with performance statistical models, 240 management changes, 214 monitoring organizational change, 208-209 decisions, 159 portfolio risk, 84 proactive changes, 163-167 risk, analytics, 8-10 reactive changes, 159-162 tradeoffs, 17-18 models, 163 mapping decisions, 96-97 KPIs, 96-97 market of one, 31 experience of one, 32 paralysis of choice, 31 Index 281 multiple approaches, appropriate optimization models, Decision responses, 172 Services, 141-143 A/B testing, 173 constraints, 144 champion-challenger testing, 174 decision variables, 144 multiple similar processes, process- iterate, 147 centric, 90 objectives, 144 solvers, 146 N optimization systems, 243-244 optimizing, predictions, 147 Netflix, 32 organizational change, 206-207 New York Department of Taxation difficulties of, 207-208 and Finance, 38-39 managing, 208-209 non trivial, characteristics of suitable reasons for, 209 decisions, 75 sponsorships, 209 continued updates, 78 organizational issues, three-legged decision is made at defined times, 75 stool, 196 domain knowledge and expertise, 76 organizations, changing, 27-28 large amounts of data, 77 orient, OODA loop, 233 large numbers of actions, 78 orthogonality, 227 policies and regulations, 76 required analysis, 77 trade-offs, 78 P-Q non-event data and event data, 92-93 paralysis of choice, 31 patterns, 216 O Decision Service integration patterns, 221-222 objectives, optimization models, 144 incremental patterns, 222 objects, executable rules, 131 strategic patterns, 221 observe, OODA loop, 233 tactical patterns, 221 OODA (Observe-Orient-Decide- pay and chase, 39 Act), 53 performance, Decision Services, 171 OODA loop (Observe, Orient, Decide, performance management, linking Act), 232 with decision management, 167 act, 234 performance management systems, 96 decide, 234 analyzing past data, 97 observe, 233 mapping decisions and KPIs, 96-97 orient, 233 root cause analysis, 98 operational databases, 247-248 personalization, 32 operational decisions, 50-51, 222 policies, reactive changes, 160-161 opportunity, predicting, 62 policies and regulations, Decision opportunity decisions, 85 Services, 125-126 optimization, refining decision- pooled data sources, 106 making approaches, 174-176 portfolio risk, managing, 84 282 Decision Management Systems pre-configured decision management repeatable decisions, 49-50 systems, 244-245 tactical decisions, 52-53 cons of, 246 test, learn, and continuously improve, pros of, 245 63-65 predicting collect and use information to improve, 65 fraud, 62 optimize over time, 66 impact of decisions, 63 support experimentation, 65 opportunity, 62 prioritizing decisions, 111-113 risk, 62 proactive changes, 163 predictions, optimizing, 147 building decision monitoring predictive analytic models, 223, 240 environments, 165-166 building, 239 capturing decision effectiveness Decision Services, 137-138 data, 164 building and testing, 139 linking performance management and deploying, 140 decision management, 167 exploring and understanding data, 138 process-centric, finding decisions, 89 preparing data, 138-139 decision points, 89 selecting techniques, 139 escalation and referral, 91 updating local exceptions, 90 champion-challenger, 180 multiple similar processes, 90 model refresh, 179 process focus, 56 self-learning, 179 process improvement, agility, 7 predictive analytics, 83 process integration, Decision Predictive Analytics Workbench, Services, 150 239-241 project goals, 111 preparing data for predictive analytic models, 138-139 R presenting data with decisions randomization, 226 together, fact-based ratios, breaking, 36-38 decisions, 230 reactive changes, 159 principles of Decision Management changes to business goals and metrics, Systems, 48 159-160 be predictive not reactive, 60-61 changes to underlying data, 161-162 predicting impact of decisions, 63 new regulations or policies, 160-161 predicting opportunity, 62 real-time responsiveness, consumer predicting risk or fraud, 62 expectations, 20-21 be transparent and agile, 57 reasons for organizational change, 209 business agility, 59-60 reducing fraud, analytics, 10-11 design transparency, 58 refactoring projects, decision execution transparency, 58-59 inventory, 213 begin with a decision in mind, 48-49 referrals, process-centric, 91 Decision Support Systems, 54-55 regulations, reactive changes, 160-161 micro decisions, 52 operational decisions, 51 Index 283 repeatability, characteristics of turning dependency networks into flow, suitable decisions, 72 119-121 consistent measurement of success, 75 scale of business operations, 23 decision is made at defined times, 73 big data, 24 defined set of actions, 74 efficiency, 24-25 same information is used each time, 73 transaction volumes, 25 repeatable decisions, 49-50 SDLC (software development repeatable management decisions, 86 lifecycle), adapting, 215-219 replication, 227 multiple data models and business resolving organizational issues, three- process management, 220 legged stool, 196 selecting business rules resources, focusing, 13-14 representation, 132-134 responses, determining appropriate self-learning, predictive analytic responses, 167-168 models, 179 assessing decision effectiveness, self-service, consumer expectations, 170-172 22-23 assessing design impact, 168-169 Sequoia Hospital, 46 comparing existing approaches, 172 service contracts, 117-119 determining if multiple new service-oriented platforms, 255 approaches are required, 172-174 BPMS (Business Process Management optimization to refine decision-making System), 257-258 approaches, 174-176 Event Processing System, 258-261 retaining, 11-13 SOA, 256 revenue, maximizing, 44-45 simulations, impact of decision- Richmond Police Department, 13-14 making approaches, 185 risk Smart Grids, 42 managing, 8-10 Smart Meters, 42 predicting, 62 smart people, 45-46 risk decisions, 82-83 SOA (service-oriented root cause analysis, metrics and architecture), 256 performance management social interactions, 26-27 systems, 98 solvers Roses, Allen, 32 optimization models, 146 rule management environments, optimization systems, 244 building, 177-179 source rules rule sheets, 132 Decision Services, 123-124 linking to, 135 S sponsorships, organizational change, 209 scaffolding, building Decision Starbucks, 32 Services, 116-117 statistical awareness, fact-based build test and learn infrastructure, decisions, 229-230 121-122 statistical models, 240 defining service contracts and test cases, 117-119 284 Decision Management Systems strategic decisions, 50-51, 222 Decision Discovery, 194-195 strategic patterns, 221 resolving organizational issues, 196 structured content, 105 starting three-way converßsations, 193 suitability, 80 time to value, decision suitable decisions, characteristics characteristics, 110 of, 72 timeliness, decision candidates for automation, 80-81 characteristics, 109 measurable business impact, 79 top-down, finding decisions, 88 non trivial, 75 traceability, 169 non trivial, analysis, 77 business rules, 136 non trivial, continued updates, 78 tradeoffs, managing, 17-18 non trivial, domain knowledge and transaction volumes, scale of big expertise, 76 business, 25 non trivial, large amounts of data, 77 transparency non trivial, large numbers of design transparency, 58 actions, 78 execution transparency, 58-59 non trivial, policies and regulations, 76 travel agencies, 44 non trivial, trade-offs, 78 trigger action, Event Processing repeatability, 72 System, 260 repeatability, consistent measurement of success, 75 U repeatability, decision is made at underwriters, 36 defined times, 73 unstructured content, 105 repeatability, defined set of actions, 74 updating predictive analytic models repeatability, same information is used champion-challenger, 180 each time, 73 model refresh, 179 synchronizing with performance self-learning, 179 management changes, decision inventory, 214 V T validating business rules, 136 validation, 82 table-driven code, 95 value range, decision tactical decisions, 50, 52-53 characteristics, 109 tactical patterns, 221 verifying business rules, 136 targeting, 11-13 versioning, Decision Services, 256 telecommunications, maximizing volume, decision characteristics, 108 assets, 41 templates, BRMS, 134 test cases, 117-119 W-X-Y-Z testing, 16-17 what-if analysis, 98 three-legged stool, 191-193 impact of decision-making aligning around business results, 195 approaches, 186 building collaboration skills, 193-194 writing executable business rules, 134 This page intentionally left blank